RuntimeError:size of tensor in conditional design

[DaHenz]

Hi, many thanks for the great and creative program. Follow the instructions, the unconditional work is smoothly finished. But the conditional design met an issue like the following: /home/sliu/anaconda3/envs/delle/bin/python draw_samples.py --input_pdb my_pdb.pdb --resample_idxs 0-25,70-80 Namespace(input_pdb='my_pdb.pdb', maxlen=60, minlen=50, model_checkpoint='checkpoints', modeldir=None, modelepoch=None, mpnnpath='checkpoints/minimpnn_state_dict.pth', num_lens=None, param=None, parampath=None, paramval=None, perlen=2, resample_idxs='0-25,70-80', steplen=5, targetdir='.', type='allatom') Warning: when sampling conditionally, the input pdb length (373 residues) is used automatically for the sampling lengths. Base directory: . Samples saved to: ./samples Model loaded from checkpoints/allatom_state_dict.pth Beginning sampling for 23-10-21-01-46-31... Traceback (most recent call last): File "draw_samples.py", line 356, in <module> main() File "draw_samples.py", line 333, in main sampling_time = draw_and_save_samples( File "draw_samples.py", line 58, in draw_and_save_samples aux = sampling.draw_allatom_samples( File "/home/sliu/Tools/protpardelle/sampling.py", line 157, in draw_allatom_samples aux = model.sample( File "/home/sliu/Tools/protpardelle/models.py", line 561, in sample gt_atom_mask = utils.atom37_mask_from_aatype(gt_aatype, seq_mask) File "/home/sliu/Tools/protpardelle/core/utils.py", line 98, in atom37_mask_from_aatype atom_mask *= seq_mask[..., None] RuntimeError: The size of tensor a (373) must match the size of tensor b (50) at non-singleton dimension 1

No matter how I change the input pdbs or the arguments, always the same error occurs. Hope anyone can help with this!

[zhkim216]

Hello DaHenz, Sorry for replying late and thank you for your interest in our model! We're currently updating the code and will fix this issue soon, maybe in this week :)

[satyabikash]

is this issue resolved ? I'm still facing the same issue ?